We propose to identify novel double strand break (DSB) producing compounds which interact with the human tumor cellular context in a manner which is distinct from existing agents. We hypothesize that by using a cell based approach we have identified an entirely new class of DSB producing compounds. For simplicity, these compounds will be referred to throughout this proposal as 'putative DSB producing' compounds. A subset of the new compounds are structurally novel and do not behave like topo I or II poisons. They may thus represent entirely new classes of DSB inducing agents. Topoisomerase poisons are successfully used in the clinic against different cancers. Unfortunately these successes are plagued by severe dose-limiting toxicities and ovelapping mechanisms of resistance. The compounds we propose to study were isolated in a genetic defect-based rather than a narrowly defined target-based screen. The selection criteria eliminate the majority of agents that directly damage DNA. This approach makes no selection based on chemistry or mechanism of action, and may have identified compounds that interact with new classes of proteins to damage DNA. We propose to carry out indicative toxicity tests in yeast and mammalian cells and on purified topoisomerases to clearly categorize the compounds into topoisomerase I poisons, topoisomerase II poisons, and mechanistically novel putative double strand break-causing agents. We will then confirm the production of double strand breaks in vivo. Finally we will identify the protein targets and other cellular factors which influence the cytotoxicity of the compounds. This will be achieved by profiling the toxicity of compounds in a panel of mammalian matched pairs of cell lines +/- defects associated with cancer, and identifying cellular factors in yeast, by overexpression and deletion studies. The yeast experiments will involve both classical genetic approaches and genome-wide analyses based on DNA microarray technology.